Algebraic Model for Agent Explicit Knowledge in Multi-agent Systems

Information security is an important aspect that should be considered during system development. Analyzing the specification of a system enables detecting flaws at early stage of a system development. An agent knowledge of the exchanged information and its nature is essential for analyzing systems. An agent can enrich its knowledge by receiving information as messages and producing new information from the existing one. We classify an agent knowledge as explicit knowledge and procedural knowledge. The explicit knowledge of an agent is related to the information that it possesses. For example in the context of a hospital software, an agent explicit knowledge would contain information about patients, drugs, and diseases. In the context of a school, the explicit knowledge of an agent would contain information about students, courses, and instructors. In the area of cryptographic protocols, the information of an agent can be its own key, the cipher used for encryption and decryption, and the identity of other agents that are involved in the protocol. Agents communicate by sending messages which are pieces of information stored in their explicit knowledge. The information an agent receives from other agents becomes a part of its explicit knowledge. The procedural knowledge involves a set of mechanisms/operators that enables an agent to obtain new information from its explicit knowledge. For example, if the explicit knowledge of an agent contains an encrypted message as well as the key and the cipher used to decrypt the message, then by using the procedural knowledge, the concealed information can be obtained. The use of the procedural knowledge to analyze cryptographic protocols can be found in Sabri and Khedri (2006; 2007b). The explicit knowledge representation is needed to analyze security related policies in multiagent systems. We summarize below some uses of the explicit knowledge: